算力
Search documents
沐曦施淑珏:国产GPU商业化加速 软硬协同跨越“好用”门槛
2 1 Shi Ji Jing Ji Bao Dao· 2026-01-23 10:55
在国产GPU四小龙接连登陆资本市场的同时,国产算力芯片也进入大规模产业化落地的关键时期。现 场,施淑珏还谈到沐曦在商业化方面的探索,以及新一年的发展规划情况。 商业化核心门槛在于从"可用"到"好用"的跨越 施淑珏表示,沐曦已构建了从硬件架构到软件栈的完整自研技术体系。凭借全建制的研发团队,沐曦得 以推出覆盖人工智能训练和推理、通用计算、图形渲染的全栈GPU产品矩阵。目前,沐曦已具备从芯 片、到板卡、服务器、超节点直至部署大规模算力集群的全链条能力。 近年来,国产GPU的落地应用情况成为业界重点关注的话题。施淑珏指出,国产GPU商业化的核心门槛 在于从"可用"到"好用"的跨越。在她看来,衡量一款GPU是否好用,除了单芯片算力,也需要多维度的 能力,沐曦产品在通用性、易用性、稳定性和规模可扩展性四个方面均有亮眼的表现。 "我们更加注重真正交付给客户在AI应用场景实战中的产品性能。"她说。 近年来,在人工智能浪潮席卷全球、算力需求呈指数级增长的背景下,国产GPU成为资本市场瞩目的焦 点。2025年12月,摩尔线程、沐曦股份的接连科创板上市,再度将市场对国产GPU的追逐推向高潮。 近日,胡润研究院在北京亦庄发布的《2 ...
沐曦施淑珏:国产GPU商业化加速,软硬协同跨越“好用”门槛
2 1 Shi Ji Jing Ji Bao Dao· 2026-01-23 10:35
21世纪经济报道记者申俊涵北京报道 近年来,在人工智能浪潮席卷全球、算力需求呈指数级增长的背景下,国产GPU成为资本市场瞩目的焦点。2025年12月,摩尔 线程、沐曦股份的接连科创板上市,再度将市场对国产GPU的追逐推向高潮。 近日,胡润研究院在北京亦庄发布的《2025胡润中国人工智能企业50强》显示,寒武纪、摩尔线程、沐曦股份成为排名前三的 人工智能企业。据悉,该榜单按照企业价值进行排名,上市公司市值按照2026年1月9日的收盘价计算,非上市公司估值参考同 行业上市公司或者根据最新一轮融资情况进行估算。 "我们更加注重真正交付给客户在AI应用场景实战中的产品性能。"她说。 据悉,沐曦是国内少数实现千卡集群大规模商业化应用的GPU供应商,正研发和推动万卡集群的落地。施淑珏认为,这考验的 是中国企业最擅长的"工程能力"。事实上,中国企业在芯片设计方面,已经能够跟国际企业进行同等对话。 其中,寒武纪凭借 6300亿元的价值,位居榜单首位。2025年上半年,寒武纪实现营收28.8亿元,同比大幅增长43倍;摩尔线程 以3100亿元的价值排名第二,2025年公司以创科创板IPO最快过会纪录——88天。2025年前三季度 ...
算力告急!“缺口”风暴下,国产AI芯片如何突围
Xin Lang Cai Jing· 2026-01-23 10:27
Core Insights - The announcement from Beijing Zhiyuan Huazhang Technology Co., Ltd. highlights a significant increase in user demand for AI models, leading to a temporary strain on computing resources, reflecting a broader trend in the AI industry [1] - The AI chip market in China is projected to exceed 1 trillion yuan by 2028, accounting for approximately 30% of the global market, emphasizing the need for high-quality, domestically controlled AI computing power to seize opportunities in the AI sector [1] Group 1: Current State of Computing Power - There is a pronounced supply-demand imbalance in computing power, particularly in China, which is more evident than in the global context [3] - Foreign companies dominate the global AI computing power market, holding nearly 70% of the market share in China by 2024, creating a substantial self-sufficiency gap for domestic production [4] - The self-sufficiency rate of AI GPUs in China has increased from less than 10% in 2020 to approximately 34% in 2024, with expectations to reach around 82% by 2027 [4] Group 2: Causes of Supply-Demand Imbalance - The supply of computing resources in China faces multiple constraints, including limitations on high-end chip imports and performance gaps between domestic and international GPU products [5] - The fragmentation of computing resources among service providers leads to low utilization rates, exacerbating the supply-demand mismatch [5] - The rapid deployment of AI applications across various industries has resulted in over 13,000 projects and more than 30,000 smart factories, significantly increasing the demand for computing power [5] Group 3: Solutions to Computing Power Challenges - To address the computing power challenges, it is essential to maximize the potential of domestic computing resources and promote their application [6] - The Chinese government has been actively implementing policies to enhance computing power development, including optimizing infrastructure and improving service levels [6] - Collaboration among stakeholders in the computing power ecosystem is crucial for achieving deep integration of models, applications, and computing resources, which will enable China to gain a competitive edge in the global AI landscape [7]
“不务正业”,中国移动下场卖保险
Tai Mei Ti A P P· 2026-01-23 10:04
Core Insights - China Mobile officially enters the insurance market, leveraging its vast user base of nearly 1 billion and extensive offline channels to disrupt the competitive landscape of the insurance industry [1][10][14] Group 1: Strategic Entry into Insurance - The entry into insurance is a culmination of over 20 years of strategic positioning, starting with the establishment of a national insurance brokerage in 2001, which allowed China Mobile to navigate regulatory frameworks and sales processes effectively [1][2] - In 2016, China Mobile became the second-largest shareholder of China Merchants Renhe Life Insurance, marking a shift from simple channel cooperation to a deeper capital partnership aimed at creating synergistic effects between channels, capital, and products [2][10] - The launch of "Home Insurance" and "Medical Insurance" in early 2026 represents a significant step towards large-scale operations, with initial trials focused on local markets to assess product fit and user acceptance [2][3] Group 2: Product Offerings and Market Response - The "Medical Insurance" product addresses the common pain point of high medical costs, offering low entry barriers and comprehensive coverage, with annual premiums ranging from 199 to 699 yuan and a maximum coverage of 1.118 million yuan [4][10] - The "Home Insurance" product combines property and accident coverage with value-added services, offering flexible pricing from free to 199 yuan per month, with a maximum coverage of 1.442 million yuan [5][6] - Initial market feedback indicates that while the products are attracting attention due to their integration with communication services, there are concerns regarding limited coverage and unclear terms [3][4] Group 3: Financial Performance and Market Context - China Mobile's financial performance remains strong, with a revenue of 1,040.76 billion yuan in 2024, a year-on-year increase of 3.12%, and a net profit of 138.4 billion yuan, reflecting robust profitability and cash flow stability [7][10] - The overall telecom industry is experiencing a slowdown in revenue growth, prompting China Mobile to seek new profit avenues through insurance, which offers stable cash flow and long-term lifecycle benefits [10][12] - The shift towards insurance is seen as a strategic move to diversify revenue streams and alleviate pressure from traditional business segments, which are nearing saturation [10][12] Group 4: Competitive Landscape and Industry Dynamics - The competitive landscape among telecom operators is evolving, with China Unicom focusing on AI and partnerships, while China Telecom is retracting from the insurance sector, highlighting differing strategic responses to market pressures [3][8] - The insurance market in China is characterized by intense competition, with significant growth potential in health and property insurance, particularly in underserved markets [11][12] - China Mobile's strategy to leverage its extensive offline presence and user trust aims to enhance accessibility and integration of insurance services into everyday life, positioning it as a "hidden guardian" in consumers' daily activities [14]
美克家居现资金链危局?控股股东质押“爆表”股份遭冻结 跨界算力能否自救
Xin Lang Cai Jing· 2026-01-23 10:03
Core Viewpoint - Meike Home is facing a potential liquidity crisis due to significant short-term debt and ongoing operational losses, compounded by the freezing of shares held by its controlling shareholder and concerns over the quality of its recent acquisition target [1][15][10]. Financial Performance - The company has experienced continuous losses since 2021, with total revenue declining from 52.75 billion in 2021 to an estimated 22.23 billion in 2025, reflecting a year-on-year decrease of 10.10% [4][18]. - The net profit attributable to shareholders has also deteriorated, with a projected loss of 2.20 billion in 2025, following a trend of negative profits since 2021 [4][18]. Debt Situation - As of the third quarter of 2025, Meike Home's short-term borrowings exceeded 18 billion, while cash on hand was only 1.16 billion, indicating severe liquidity pressure [7][21]. - The controlling shareholder's shares have been fully pledged, raising concerns about potential margin calls and further financial instability [10][23]. Operational Challenges - The company has seen a significant reduction in its domestic retail operations, with the number of stores dropping from over 470 in 2022 to 327 in 2024, a decline of over 30% [5][19]. - Reports indicate that Meike Home has been delaying salary payments to over 400 employees and owes substantial amounts to suppliers, leading to collective employee actions [5][19]. Acquisition and Strategic Shift - Meike Home has announced plans to acquire 100% of Wande Technology, a company specializing in high-speed copper cables, as part of a strategic shift towards the computing power sector [9][25]. - The financial performance of Wande Technology shows modest revenue growth, with 2023 revenue at 45.98 million and a net profit of 3.98 million, raising questions about the acquisition's potential to enhance Meike Home's profitability [12][26]. Market Reaction - Following the announcement of the acquisition, Meike Home's stock experienced significant volatility, with a near 50% increase followed by a sharp decline, indicating market uncertainty regarding the company's future [12][26].
券商业绩预喜股价却滞涨,业内分析三大背离成因
Huan Qiu Wang· 2026-01-23 07:30
针对这一背离现象,业内分析认为主要原因有三点。首先,券商行情往往具有"提前量"特征。早在2024年第四季度A股启动 上涨行情时,券商板块已抢先大涨,提前透支了2025年的业绩增长利好。当正式财报落地时,市场资金倾向于获利了结, 进入"卖事实"阶段,导致业绩增长难以推动股价进一步上行。 其次,资金虹吸效应显著。2025年市场呈现极致的结构分化,人工智能、半导体、算力等科技赛道因具备高成长性和广阔 的想象空间,持续吸引增量资金。相比之下,券商股被市场视为传统周期品种,缺乏短期的爆发性叙事,导致资金吸引力 下降,"牛市旗手"的光环有所褪色。 【环球网财经综合报道】尽管上市券商2025年业绩预喜,净利润大幅增长,但其股价表现却未能同步跟进,呈现出明显 的"滞涨"态势。近期披露的业绩预告显示,多家券商营收与净利双增,但在二级市场上,券商板块去年整体跑输大盘,这 种"业绩热、股价冷"的反差引发了市场的广泛关注。 回顾2025年全年,A股市场震荡走高,结构性行情突出。科技赛道表现抢眼,主要指数涨幅显著,而券商板块表现相对低 迷,未能跑赢沪深300等大盘指数。然而,从基本面来看,券商行业却展现出强劲的增长韧性。数据显示,上市 ...
国产AI算力生态持续完善,产业资本化趋势加速,港股通互联网ETF易方达(513040)等产品受市场关注
Mei Ri Jing Ji Xin Wen· 2026-01-23 06:31
有分析认为,从更广泛的板块视角看,这一动向反映出港股互联网龙头正加速向"模型—算力—应用— 生态"一体化体系延伸,AI不再仅停留在应用层叙事,而是逐步向基础设施与核心技术能力下沉。在当 前估值仍处于相对合理区间的背景下,围绕AI底层能力的显性化与资产化,有望推动港股科技板块迎 来新一轮价值重估。 分别跟踪以上指数的港股通互联网ETF易方达(513040)、恒生科技ETF易方达(513010)受到资金青 睐,今年以来均获约15亿元净流入,可助力投资者一键打包行业龙头企业。 1月23日午后,港股科技板块涨幅有所扩大。截至14:00,恒生科技指数上涨0.4%、中证港股通互联网指 数上涨0.8%。 近日,市场关注到阿里巴巴正推进旗下AI芯片业务平头哥的独立化运作,并拟在完成内部重组后探索 独立上市的可能性。该事件展现了阿里在AI底层技术布局上的重要进展,也为其长期投入的芯片与算 力资产提供了更清晰的价值实现路径。AI芯片业务叠加阿里在大模型与云计算领域的应用场景,使其 在算力效率与规模化落地方面具备差异化优势。 恒生科技指数由港股上市公司中与科技主题高度相关的、市值最大的30只股票组成,聚焦半导体、机器 人、软件、互 ...
创业板人工智能ETF华宝(159363)受益“算力+AI应用”,一举成双创赛道规模最大、流动性最佳人工智能ETF!
Sou Hu Cai Jing· 2026-01-23 03:02
Core Insights - The article highlights the significant growth and market interest in the Huabao AI ETF (159363), which has reached a scale of 6.303 billion yuan, making it the largest and most liquid AI-themed ETF in the dual innovation sector as of January 22, 2026 [1][4][5]. Fund Performance - The Huabao AI ETF experienced a rapid increase in scale, growing from over 5 billion yuan to over 6 billion yuan within just three trading days (January 14-19, 2026) [1]. - As of January 22, 2026, the ETF's average daily trading volume over the past six months exceeded 800 million yuan, indicating strong market activity [1][4]. Market Position - The Huabao AI ETF has surpassed 24 other similar AI-themed ETFs targeting the ChiNext and Sci-Tech Innovation Board, establishing itself as the "new king" in the dual innovation sector [1][4]. - It is the first ETF in the market to track the "ChiNext AI Index," which reflects the price changes of listed companies related to the AI theme on the ChiNext board [6][7]. Investment Strategy - The ETF is strategically positioned to capture the growth of AI commercialization, with approximately 60% of its portfolio allocated to computing power (primarily optical modules) and 40% to AI applications, effectively covering the entire AI industry chain [7][8]. - The fund manager emphasizes the ETF's dual focus on computing power and AI applications, aiming to benefit from the ongoing growth in AI commercialization [7]. Broader ETF Matrix - Huabao Fund has developed a comprehensive "AI+" ETF investment matrix, which includes various products that cover the entire AI industry chain, such as the Sci-Tech AI ETF and other thematic ETFs focused on financial technology and big data [8][9]. - The matrix also includes ETFs targeting specific applications of AI, such as healthcare and smart driving, indicating a strategic diversification in investment offerings [10][11].
双创赛道规模最大人工智能ETF!创业板人工智能ETF华宝(159363)规模跃升至63.03亿元
Xin Lang Cai Jing· 2026-01-23 02:40
Core Insights - The Huabao Entrepreneurial Board AI ETF (159363) has reached a scale of 6.303 billion yuan, becoming the largest and most liquid AI ETF in the dual innovation track as of January 22, 2026 [1][16] - The ETF's scale increased from over 5 billion to over 6 billion yuan within just three trading days, highlighting its strong market appeal [1][16] - The ETF is the first in the market to track the "Entrepreneurial Board AI Index," which reflects the stock price changes of companies related to the AI theme [5][20] Fund Performance - As of January 22, 2026, the ETF has a daily average trading volume exceeding 800 million yuan over the past six months, indicating its high liquidity [1][16] - On January 22, the ETF's price surged by 3.03%, with a net subscription of 22 million units, and it has seen an increase of nearly 2 billion yuan in the past ten days [5][20] Investment Strategy - The ETF's manager believes in a dual-line layout of computing power and AI applications, covering the entire AI industry chain from infrastructure to deep integration with various sectors [6][21] - Approximately 60% of the ETF's holdings are allocated to computing power (mainly optical modules), while about 40% are invested in AI applications, positioning it to benefit from the commercialization of AI [6][21] Product Matrix - Huabao Fund has established a comprehensive "AI+" ETF investment matrix, including various products that cover the entire AI industry chain [8][23] - The matrix includes the Sci-Tech Innovation AI ETF and other thematic ETFs focusing on financial technology, big data, and AI applications in various sectors [8][23] Market Trends - The AI sector is transitioning from concept to industrialization, with application scenario ETFs expected to witness significant potential [10][25] - The healthcare sector is identified as a crucial area for AI applications, with potential impacts on clinical diagnosis, medical software upgrades, and drug development [10][25]
北京:对能效超标的数据中心正式征收差别电价
Xin Hua She· 2026-01-23 02:15
新华社北京1月22日电,在22日举行的北京市经济和信息化局2026年一季度新闻发布会上,北京市经济 和信息化局数字产业处副处长张金瑞介绍,2026至2027年北京市将持续加大算力资源供给力度,计划到 2027年累计建成算力规模到20万P左右。 北京将优化打造"银河算廊"工程,构建跨域协同算力供给体系。具体而言,扎实建设京内算力支撑节 点,按照靠近产业集聚区、靠近新能源富集区、靠近山区的布局原则,在延庆-门头沟-房山、昌平- 海淀、亦庄-通州等区域建设超级节点和行业节点。积极共建环京算力供给走廊,引导电信运营商、头 部互联网企业和IDC企业在河北张家口、廊坊,天津武清等地建设大规模算力集群,重点服务人工智能 产业发展。加快拓展西部算力供给走廊,支持头部互联网企业和电信运营商在内蒙古、宁夏和青海等西 部地区建设大规模训练算力集群,依托算力互联网汇聚优质算力资源,支撑基础模型国际争先。 2026年,北京市将推进40家存量数据中心绿色低碳改造,对能效超标的数据中心正式征收差别电价;推 动数据中心进入电力交易市场,不断扩大数据中心绿电消纳比例;探索实施数据中心余热利用、算电协 同等示范项目,全面推动存量数据中心提质升级 ...